- Docente: Filippo Ferrari
- Credits: 6
- SSD: M-FIL/05
- Language: Italian
- Teaching Mode: Traditional lectures
- Campus: Bologna
- Corso: First cycle degree programme (L) in Communication Sciences (cod. 5975)
-
from Nov 10, 2025 to Dec 18, 2025
Learning outcomes
In this course we will examine, from a philosophical perspective, some biases (especially biases related to cultural, social, and gender issues) which affect our web searches as well as our interactions with language models based on artificial intelligence like chatGPT. We will do so by using conceptual tools from epistemology and the philosophy of language. The course aims at achieving the following objectives: (i) to provide a basic knowledge of how the search algorithms and the semantic web work; (ii) to analyse the most common biases affecting web searches, with special focus on gender biases; (iii) to provide a philosophical model of the epistemic mechanism underlying these biases by means of the conceptual tools provided by analytic and social epistemology; (iv) to foster students' awareness of the problematic aspects related to biases in web searches and in interactions with artificial intelligence and to equip them with some tools to limit the negative effects of such biases.
Course contents
This course examines ethical considerations surrounding algorithms and data-driven machine learning technologies, highlighting their societal implications.
By means of knowledge and methods from analytic epistemology and philosophy of language, students will explore key issues such as bias both in general and as applied to algorithms, fairness, and epistemic injustice, particularly how machine learning systems can unintentionally reproduce or amplify harmful social biases related to race, gender, and other categories.
The course introduces fundamental concepts like fairness, bias, epistemic injustice, methods for detecting and mitigating bias. Additional topics may include analyzing sources of bias, evaluating fairness metrics, and understanding the limitations of current fairness frameworks, including the so-called "unfair fairness". The goal is to equip students with core philosophical skills and perspectives necessary to critically assess and responsibly engage with digital technologies.
Readings/Bibliography
Fundamental readings (selection from)
- O'Neil, Cathy (2016) Weapons of Math Destruction, New York: Crown.
- Umoja Noble, Safiya (2018) Algorithms of Oppression, New York:New York University Press.
- Floridi, L. (2023) The Ethics of Artificial Intelligence, Oxford: Oxford University Press.
Additional readings (excerpts)
- Allen, J. P., & Smit, H. (2020). Crowdsourcing moral machines. Science, 367(6485), 1418–1419. https://doi.org/10.1126/science.367.6485.1418
- Belot, G. (2016). A Defence of the Ideal vs. Non-Ideal Distinction. Mind, 125(500), 1105–1142. https://doi.org/10.1093/mind/fzv185
- Bradley, S. (2024). Explanation and understanding. Philosophical Studies, 181(4). https://doi.org/10.1007/s11098-024-02273-w
- Brown, M. J. (2022). Responsible innovation and algorithmic bias. Res Publica, 28(4), 569–589. https://doi.org/10.1007/s11158-022-09546-3
- Caliskan, A., Bryson, J. J., & Narayanan, A. (2017). Semantics derived automatically from language corpora contain human-like biases. Science, 356(6334), 183–186. https://doi.org/10.1126/science.aal4230
- Eva, B. (2022). Algorithmic fairness and base rate tracking. Philosophy & Public Affairs, 50(2), 239–268. https://doi.org/10.1111/papa.12211
- Fazelpour, S., & Danks, D. (2020). Algorithmic bias: Senses, sources, solutions. Philosophy Compass, 16(8), e12760. https://doi.org/10.1111/phc3.12760
- Fleisher, W. (2020). Rational endorsement. Ergo, 6(15). https://doi.org/10.3998/ergo.12405314.0006.015
- Frith, C. D. (2017). Mind-reading machines. Science, 356(6334), 133–134. https://doi.org/10.1126/science.aan0649
- Hempel, C., & Knüsel, C. (2021). The Bias Blind Spot: A Review. SSRN. http://dx.doi.org/10.2139/ssrn.3974963
- Himmelspach, S. (2023). Explainability as a requirement for trustworthy AI. Philosophy & Technology, 36(4). https://doi.org/10.1007/s13347-023-00679-8
- Holm, S. (2022). The Fairness in Algorithmic Fairness. Preprint.
- Hummel, P. (2025). Algorithmic fairness as an inconsistent concept. American Philosophical Quarterly, 62(1).
- Jansson, L. (2020). Epistemic risk and epistemic injustice. Synthese, 198, 11645–11667. https://doi.org/10.1007/s11229-020-02696-y
- Kappel, K. (2023). Misleading higher-order evidence. International Journal of Philosophical Studies, 31(1), 25–45. https://doi.org/10.1163/17455243-20234372
- Noble, S. U. (2016). Algorithms of oppression. Big Data & Society, 3(2). https://doi.org/10.1177/2053951716649398
- Pedersen, N. J. L. L. (2024). Epistemic injustice and AI. Episteme. https://doi.org/10.1017/epi.2024.11
- Schulz, K. (2023). Epistemic trespassing and expertise. Philosophical Studies, 180(9), 2603–2621. https://doi.org/10.1007/s11098-023-02095-2
- Steel, D. (2018). Mechanisms and the evidence hierarchy. Philosophy & Public Affairs, 46(2), 105–135. https://doi.org/10.1111/papa.12189
- Steele, K. (2022). Should we be Bayesians about bias? Canadian Journal of Philosophy, 52(1), 64–81. https://doi.org/10.1017/can.2022.3
- Watson, D. (2024). Algorithms and moral responsibility. Law and Philosophy, 43, 365–390. https://doi.org/10.1007/s10982-024-09505-4
Teaching methods
Lectures supported by handouts and/or slides and supplementary materials (videos, podcasts, etc.); group work and in-class discussions designed to promote horizontal/peer learning and thematic in-depth exploration of core topics.
Assessment methods
For those who will attend the course, the overall assessment consists in the following:
-
A brief essay (between 2500 and 3500 words) in which the student critically discusses one of the topics dealt with during the course (weight: 60% of the total mark).
-
A short oral exam aimed at discussing the essay (and to prove that it has been written by the student) as well as the student's knowledge and understanding of some basic concepts dealt with during the course (weight: 40% of the total mark).
For those who will not attend the course, the overall assessment consists in the following:
-
A written essay (between 4000 and 5000 words) in which the student critically discusses one of the topics dealt with during the course (weight: 50% of the total mark).
-
A (longer) oral exam aimed at discussing the essay as well as the student's knowledge and understanding of some basic concepts dealt with during the course (weight: 50% of the total mark).
In order to be regarded as attending the course the student has to take part in at least 60% of the classes (i.e. 9 lectures out of 15).
Evaluation criteria
Concerning the essay, the basic criteria for the evaluation are: (i) whether and to what extent the essay shows an adequate knowledge and understanding of the main topics and arguments dealt with in the essay; (ii) whether the essay is adequately structured (as indicated by the guidelines made available during the course); (iii) clarity of exposition and argumentative rigor. Further criteria which, if present, may increase the evaluation are: (iv) some originality in either the content or the argumentative structure; (v) ability to critically assess in an autonomous manner the contents and arguments dealt with in the essay; (vi) ability to connect profitably the topic dealt with in the essay with some of the other topics discussed during the course; (vii) whether the student is able to autonomously perform bibliographical and thematic searches on the topic of the essay.
Concerning the oral part of the exam, the criteria for the evaluation are: (i) the extent to which the student knows and understands in a critical manner the topic of the essay, also in relation to the broader context of the course; (ii) the extent to which the student knows and understands the main topics discussed during the course—other than those discussed in the essay (this will weigh more for those students who didn’t attend the course).
Assessment Grid
30 (cum laude) — Excellent overall performance which demonstrates a solid knowledge as well as a deep and critical understanding of the topics dealt with during the course
30 — Very good overall performance which demonstrates solid knowledge and a very good understanding of the topics dealt with during the course
29-27 — Good overall performance which demonstrates a good knowledge and understanding of the topics dealt with during the course
26-24 — Fair overall performance which demonstrates adequate knowledge and understanding, but with detectable lacunae, of the topics dealt with during the course
23-20 – Sufficient overall performance which demonstrates barely adequate knowledge and understanding, with important lacunae, of the topics dealt with during the course
19-18 — Barely sufficient overall performance which demonstrates a rather superficial knowledge and understanding of the topics dealt with during the course
17 or less – Insufficient overall performance which demonstrates significant failures of understanding as well as absence of knowledge of significant parts of the topics dealt with during the course. Exam failed.
Students with disabilities and Specific Learning Disorders (SLD)
Students with disabilities or Specific Learning Disorders are entitled to special adjustments according to their condition, subject to assessment by the University Service for Students with Disabilities and SLD. Please do not contact teachers or Department staff, but make an appointment with the Service. The Service will then determine what adjustments are specifically appropriate, and get in touch with the teacher. For more information, please visit the page of the University of Bologna DSA service [https://site.unibo.it/studenti-con-disabilita-e-dsa/en/for-students].
Teaching tools
Handouts, slides, Virtuale, Wooclap, videos and podcasts by experts on core course topics.
Office hours
See the website of Filippo Ferrari